Back to results
US 12,718,703 B2Granted
Device and method for recommending educational content
AssigneeSocra AI Inc.
InventorHyun Bin Loh
Life of this patent
Estimating the term…
Abstract
Provided are a device and method for recommending educational content. The method includes acquiring a target user's learning data which includes log data including question data related to a question previously answered by the target user and answer data related to the target user's answer to the question, acquiring a question database including at least one candidate question, calculating the target user's predicted correct answer rate information for the candidate question on the basis of the candidate question and the learning data, acquiring the target user's ability information related to at least some of the log data, and determining recommendation content on the basis of the target user's ability information.
Claims9
A computer-implemented method of recommending educational content by a device for analyzing learning data of a user, the device comprising a memory storing computer-readable instructions and a processor configured to execute the instructions to perform the method comprising:
- acquiring, by the processor, learning data of a target user which includes log data including question data related to a question previously answered by the target user and answer data related to an answer of the target user to the question;
- acquiring, by the processor, a question database including at least one candidate question and average correct answer rate information of reference users for the candidate question;
- acquiring, by the processor, predicted correct answer rate information of the target user for the candidate question by executing an artificial neural network model on the basis of the candidate question and the learning data, the artificial neural network model consisting of a transformer-based artificial neural network, wherein the acquiring of the predicted correct answer rate information of the target user for the candidate question comprises calculating the predicted correct answer rate information of the target user by inputting candidate question information to an encoder of the transformer-based artificial neural network and inputting the learning data to a decoder of the transformer-based artificial neural network;
- acquiring, by the processor, ability information of the target user related to at least some of the log data on the basis of the predicted correct answer rate information and the average correct answer rate information of the reference users for the candidate question, wherein the ability information is related to at least one of first ability information representing a maximum learning ability of the target user, second ability information representing reasoning power of the target user, third ability information representing a logical ability of the target user, and fourth ability information representing an application ability of the target user, and wherein the acquiring of the ability information comprises acquiring the first ability information representing the maximum learning ability of the target user;
- generating, by the processor, an educational content set by associating the first ability information of the target user with data at a specific point of the log data included in the learning data;
- and
- determining, by the processor, the educational content set as recommendation content,
- wherein the acquiring of the first ability information of the target user comprises:
- (i) predicting, by the processor, a learning ability value of the target user by using a trained neural network model,
- (ii) generating, by the processor, growth curve data indicating a change in the predicted learning ability value of the target user over time on the basis of the learning data and the predicted correct answer rate information,
- (iii) performing, by the processor, a derivative operation on the growth curve data to acquire rate-of-change information of the growth curve data,
- (iv) acquiring, by the processor, target rate-of-change information having a rate of change smaller than or equal to a predetermined threshold rate of change at a time point on the growth curve data on the basis of the rate-of-change information, and
- (v) acquiring, by the processor, the predicted learning ability value at the time point on the growth curve data, as the first ability information representing the maximum learning ability of the target user corresponding to the target rate-of-change information.
10 more elements
A non-transitory computer-readable recording medium in which a computer program executed by a computer is recorded, the computer program comprising:
- acquiring learning data of a target user which includes log data including question data related to a question previously answered by the target user and answer data related to an answer of the target user to the question;
- acquiring a question database including at least one candidate question and average correct answer rate information of reference users for the candidate question;
- acquiring predicted correct answer rate information of the target user for the candidate question by executing an artificial neural network model on the basis of the candidate question and the learning data, the artificial neural network model consisting of a transformer-based artificial neural network, wherein the acquiring of the predicted correct answer rate information of the target user for the candidate question comprises calculating the predicted correct answer rate information of the target user by inputting candidate question information to an encoder of the transformer-based artificial neural network and inputting the learning data toa decoder of the transformer-based artificial neural network;
- acquiring ability information of the target user related to at least some of the log data on the basis of the predicted correct answer rate information and the average correct answer rate information of the reference users for the candidate question, wherein the ability information is related to at least one of first ability information representing a maximum learning ability of the target user, second ability information representing reasoning power of the target user, third ability information representing a logical ability of the target user, and fourth ability information representing an application ability of the target user, and wherein the acquiring of the ability information comprises acquiring the first ability information representing the maximum learning ability of the target user;
- generating an educational content set by associating the first ability information of the target user with data at a specific point of the log data included in the learning data;
- and
- determining the educational content set as recommendation content,
- wherein the acquiring of the first ability information of the target user comprises:
- (i) predicting a learning ability value of the target user by using a trained neural network model,
- (ii) generating growth curve data indicating a change in the predicted learning ability value of the target user over time on the basis of the learning data and the predicted correct answer rate information,
- (iii) performing a derivative operation on the growth curve data to acquire rate-of-change information of the growth curve,
- (iv) acquiring target rate-of-change information having a rate of change smaller than or equal to a predetermined threshold rate of change at a time point on the growth curve data on the basis of the rate-of-change information, and
- (v) acquiring the predicted learning ability value at the time point on the growth curve data as the first ability information representing the maximum learning ability of the target use corresponding to the target rate-of-change information.
10 more elements
A device for recommending educational content by receiving learning data of a user from an external user terminal, the device comprising:
- a transceiver configured to communicate with the user terminal;
- and
- a controller configured to acquire learning data of a target user through the transceiver and calculate learning ability of the target user on the basis of the learning data, wherein the controller is further configured to:
- acquire the learning data of the target user which includes log data including question data related to a question previously answered by the target user and answer data related to an answer of the target user to the question, acquire a question database including at least one candidate question and average correct answer rate information of reference users for the candidate question,
- acquire predicted correct answer rate information of the target user for the candidate question by executing an artificial neural network model on the basis of the candidate question and the learning data, the artificial neural network model consisting of a transformer-based artificial neural network, wherein the controller is further configured to calculate the predicted correct answer rate information of the target user by inputting candidate question information to an encoder of the transformer-based artificial neural network and inputting the learning data to a decoder of the transformer-based artificial neural network,
- acquire ability information of the target user related to at least some of the log data on the basis of the predicted correct answer rate information and the average correct answer rate information of the reference users for the candidate question, wherein the ability information is related to at least one of first ability information representing a maximum learning ability of the target user, second ability information representing reasoning power of the target user, third ability information representing a logical ability of the target user, and fourth ability information representing an application ability of the target user, and wherein the acquiring of the ability information comprises acquiring the first ability information representing the maximum learning ability of the target user,
- generate an educational content set by associating the first ability information of the target user with data at a specific point of the log data included in the learning data, and determine the educational content set as recommendation content,
- wherein the controller is further configured to:
- (i) predict a learning ability value of the target user by using a trained neural network model,
- (ii) generate growth curve data indicating a change in the predicted learning ability value of the target user over time on the basis of the learning data and the predicted correct answer rate information,
- (iii) perform a derivative operation on the growth curve data to acquire rate-of-change information of the growth curve data,
- (iv) acquire target rate-of-change information having a rate of change smaller than or equal to a predetermined threshold rate of change at a time point on the growth curve data on the basis of the rate-of-change information, and
- (v) acquire the predicted learning ability value at the time point on the growth curve data as the first ability information representing the maximum learning ability of the target user corresponding to the target rate-of-change information.
10 more elements
Dependencies are inferred from the claim text.